Use of Secondary Slags in CompletelyRecyclable Concrete

Mieke De Schepper1; Pieter Verlé2; Isabel Van Driessche3; and Nele De Belie4

Abstract: A completely recyclable concrete (CRC) is designed to have a chemical composition equivalent to the one of general rawmaterials for cement production. By doing so, this CRC can be used at the end of its service life in the cement clinkering process withoutneed for ingredient adjustments, and with improvement of the resource efficiency of cement and concrete production. Copper slag isinteresting as an iron source for the production of such a CRC and can be added to concrete, either as alternative binder or as aggregate.By isothermal calorimetry and compressive strength tests it was found that the addition of copper slag as cement replacement is of minorinterest. But a study toward the compressive strength and durability of concrete with copper slag as aggregate replacement had promisingresults. The performance of these concretes was comparable with or even better than the reference concrete, regarding strength and mostdurability aspects such as porosity and permeability, and resistance against carbonation and chloride ingress. Only the resistance to freeze-thaw attack with deicing agents was inferior. DOI: 10.1061/(ASCE)MT.1943-5533.0001133. © 2014 American Society of Civil Engineers.

Author keywords: Concrete; Recycling; Slag; Compressive strength; Durability.

Introduction

In today’s society, concrete is a popular building material: it isstrong, gives flexibility in design, and comes with a relativelylow cost. However, like all materials, it also comes with an envi-ronmental cost. About 42% (Kogel et al. 2006) of the 15 billiontons (Langer et al. 2004) of aggregates produced each year are usedin concrete, of which only 8% are recycled aggregates (Arab 2006).This use of enormous amounts of natural raw materials is the firstenvironmental problem of the concrete industry. Indeed, at the endof a concrete’s life cycle, there is the waste production: 12–21% ofthe total waste generated in the European Union is concretedemolition waste (De Belie and Robeyst 2007; C. Fischer andM. Werge, “EU as a recycling society,” ETC/SCP working paper2/2009). These numbers are the motivation for many researchers towork toward a more sustainable construction by resource efficiencyand recycling.

To optimize product-recycling opportunities, the Cradle-to-Cradle (C2C) concept was developed (McDonough and Braungart2002). In C2C production, all material inputs and outputs are eitherseen as biological nutrients or as technical resources. Biological

nutrients can be composted or consumed whereas technical resour-ces can be recycled or reused without loss of quality.

Applying this idea to the production of concrete, a completelyrecyclable concrete (CRC) was designed (De Schepper et al. 2013;Tamura et al. 2004). In order to make CRC a valuable technicalresource for cement production, without need for ingredient adjust-ments, the concrete mixture should be chemically equivalent toa traditional cement raw meal. The primary ingredient for CRCis limestone aggregate as a source for CaO, the main ingredientfor Portland cement production. Besides natural rawmaterials, suchas porphyry aggregate (SiO2 and Al2O3 source), industrial by-products like fly ash (SiO2 and Al2O3 source), and copper slag(Fe2O3 source) are preferably used for CRC production. The hy-dration of CRC cement was found comparable with the one ofOrdinary Portland Cement (De Schepper et al. 2014).

Copper slag is a by-product obtained during matte smelting andrefining of copper. Today, common management options are recy-cling, recovering of metal and production of value-added products(e.g., abrasive tools, roofing granules, cutting tools, tiles, glass,road-base construction, rail-road ballast, and asphalt pavements;Shi et al. 2008). Still, huge amounts are disposed in dumps orstockpiles. One of the greatest potential applications is in cementand concrete (Shi et al. 2008). In cement production, it can be usedas an iron adjusting material or as partial cement replacement(Al-Jabri et al. 2006; Goni et al. 1994; Gorai et al. 2003; Najimiet al. 2011; Shi et al. 2008; Taha et al. 2004, 2007; Tixier et al.1997; Zain et al. 2004). Regarding the latter, a major concern isthe leaching of heavy metals; however, studies have shown that thisis within regulatory limits. Copper slags can also be used as fineand coarse aggregates in concrete production (Al-Jabri et al. 2009a,b, 2011; Gorai et al. 2003; Khanzadi and Behnood 2009;Shi et al. 2008; Thomas et al. 2012; Wu et al. 2010a, b). The densityof this copper slag is higher compared with natural aggregates andvaries with the Fe2O3 content between 3.16 and 3.87 g=cm3

according to Shi et al. (2008). Furthermore the water absorptionis very low [0.15–0.55% according to Shi et al. (2008)]and also the glass-like smooth surface has implications towardthe design of concrete mixtures with copper slag. Copper slag alsohas, however, favorable mechanical properties such as excellent

1Ph.D. Candidate, Magnel Laboratory for Concrete Research, GhentUniv., Technologiepark Zwijnaarde 904, 9052 Ghent, Belgium (corre-sponding author). E-mail: midschep.deschepper@ugent.be

2Engineer, Master of Science in Civil Engineering, Magnel Laboratoryfor Concrete Research, Ghent Univ., Technologiepark Zwijnaarde 904,9052 Ghent, Belgium.

3Professor, Sol-gel Centre for Research on Inorganic Powders and Thinfilms Synthesis (SCRiPTS), Ghent Univ., Krijgslaan 281- S3, 9000 Ghent,Belgium.

4Professor, Magnel Laboratory for Concrete Research, Ghent Univ.,Technologiepark Zwijnaarde 904, 9052 Ghent, Belgium. E-mail: nele.debelie@ugent.be

Note. This manuscript was submitted on January 27, 2014; approved onMay 21, 2014; published online on August 11, 2014. Discussion periodopen until January 11, 2015; separate discussions must be submitted forindividual papers. This paper is part of the Journal of Materials in CivilEngineering, © ASCE, ISSN 0899-1561/04014177(9)/$25.00.

© ASCE 04014177-1 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

soundness, good abrasion resistance, and good stability (Shiet al. 2008).

This paper focuses on the use of copper slag in CRC production.Because of its lower melting point compared to iron ore, a lowercalcination temperature within the cement manufacturing processshould be feasible (Shi et al. 2008). This creates additional benefitsfor the environment: a lower energy consumption and reduced CO2

emissions, which makes this waste product even more interesting tobe used in CRC. The question remains to how this copper slag ispreferably used in CRC. The effect of partial cement replacementby copper slag was studied on cement pastes and mortars. Theinfluence of replacing aggregates by copper slag was studied formortars and concrete.

Experimental Procedure

Materials

The copper slag used in this research is a secondary slag from aBelgian copper, lead, tin, and nickel producer who uses waste prod-ucts (e.g., old copper wires, dust from the tin production, old tele-visions and computers, and radiators from cars) as raw materials.On the one hand, a quickly cooled granulated copper slag (QCS)was used for the replacement of the fine aggregates (= sand) inconcrete and as a material for cement replacement. A slowly cooledbroken copper slag (SCS) was utilized to replace the coarse aggre-gates. Before using the copper slag as a substitution for cement,QCS was milled 3 times during 4 min at 300 rpm in a planetaryball mill (mQCS). As a cementitious binder, a CEM I 52.5 N wasused throughout all experiments. An overview of the chemical andmineralogical composition of the cement and copper slag is givenin Tables 1 and 2. In Table 2 and Fig. 1, it is seen that, as expected,the quickly cooled granulated copper slag QCS has a higher otheramorphous fraction in comparison with the slowly cooled brokencopper slag SCS.

The effect of using copper slag as cement replacement wastested on cement pastes (isothermal calorimetry) and mortars (com-pressive strength). Cement pastes were produced with an increasingcopper slag content from 0 to 60 wt% in steps of 10 wt%. For eachpaste 10 g of powder (n g mQCS with (10-n) g CEM I 52.5 N) wasmixed with 4 g of water. The reference composition of the mortarswas a standard mortar 6∶2∶1 (sand:cement:water), for which thecopper slag content replacing the cement was increasing in steps

of 20 wt% from 0 to 60 wt%. A CEN sand in accordance withNBN EN 196-1 (NBN 2005) was used to produce these mortars.

For evaluation of the aggregate replacement by copper slag,mortars (for compressive strength tests) and concrete (for compres-sive strength and durability evaluation) were made. Because of thehigher density of copper slag compared with the one of natural ag-gregates, the replacement levels are volume-based, and the densityof the concrete will increase. The copper slag used in this study hada density of 2.99 g=cm3. Standard mortars were produced, and thesand was replaced by QCS from 0 to 100 vol% in steps of 20 vol%.Based on these results, the replacement levels for the concrete werefixed, namely 20 and 40 vol% of the total aggregates volume. Onthe one hand, the fine aggregates were replaced by QCS (QCS–20and QCS-40). On the other hand, the coarse aggregates were sub-stituted by SCS (SCS-20 and SCS-40). An overview of the concretemixtures is given in Table 3. The particle size distribution of mostraw materials is given in Fig. 2. All mortar and concrete specimenswere stored in a climate chamber at 20� 2°C and RH > 95% for28 days before testing, unless specified otherwise.

Table 1. Chemical Composition (wt%) of the Used Copper Slag andCement (By XRF Analysis)

Oxide QCS SCS CEM I 52.5 N

CaO 7.09 5.18 63.12SiO2 25.87 29.33 18.73Al2O3 5.92 3.23 4.94Fe2O3 45.45 50.72 3.99MgO 0.81 0.47 1.02Na2O 0.80 1.99 0.41K2O 0.19 — 0.77SO3 0.37 0.24 3.07P2O5 0.82 0.41 —TiO2 0.31 0.10 —ZnO 8.8 5.43 —MnO 0.74 0.41 —Cr2O3 0.73 0.36 —CuO 0.35 — —PbO 0.36 — —

Table 2. Mineralogical Composition (wt%) of the Used Copper Slag andCement (By XRD Analysis)

Mineral QCS SCS CEM I 52.5 N

Alite — — 53.6Belite — — 17.0Ferrite — — 11.2Aluminate — — 6.2Gypsum — — 0.8Bassanite — — 0.9Anhydrite — — 2.1Aphthitalite — — 0.3Arcanite — — 0.5Syngenite — — 0.7Zincite 0.4 0.6 —Hematite — 0.8 —Wuestite 1.6 0.4 —Fayalite 4.1 45.5 —Hercynite 4.0 3.9 —Other 89.9 48.8 6.8

Fig. 1. Observed XRD patterns of QCS and SCS between 15 and 50°2θ; the large bump in the 20–40° 2θ range in case of QCS results fromthe scatter from the amorphous fraction (Corundum was added as in-ternal standard for quantification of the other phases)

© ASCE 04014177-2 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Methods

Isothermal Calorimetry of Cement PastesThe tested powder (cement and copper slag) was mounted in a glassvial. After the addition of water, the cement pastes were mixedmanually for 20 s in the glass vial. Subsequently, it was immedi-ately placed in an isothermal calorimeter (TAM AIR) at 20°C tomeasure the hydration heat during the first 7 days of hydration.The measurement started within 30 s after addition of water. Be-cause mixing occurred outside the calorimeter, the first hydrationpeak could not be registered entirely. Because this peak onlyamounts to a few percent of the total heat liberated (De Schutter1999), it will not be considered in further analysis. To avoid sig-nificant differences between the paste and the isothermal environ-ment, the components were kept at a temperature close to themeasurement temperature before mixing.

Compressive Strength of Mortars and ConcreteCompressive strength tests on mortar specimens (160 × 40×40 mm3) were performed according to NBN EN 196-1 (NBN2005) at the age of 7 and 28 days. The concrete cubes (150 × 150×150 mm3) were tested for their compressive strength at the age of 2,7, 28, and 84 days according to NBN EN 12390-1 (NBN 2001a).

Durability of Concrete

Open Porosity. The open porosity of the concrete specimens(drilled cores d ¼ 100 mm h ¼ 50 mm) was determined usingthe vacuum absorption test described in NBN B 05-201 (NBN1976). After drying and weighing the samples (wd–kg), they wereput under vacuum (2.7 Pa) for 3 h. Subsequently, the tank was filledwith water, and the samples were stored under water for 24 h.

Finally the mass of the specimens was determined in air(wa − kg) and water (ww–kg). The porosity (φ − vol%) of the sam-ple could then be determined by calculating the volume of the poresfilled with water

φ ¼ Vwater

Vconcrete¼

wa−wdρwaterwd

ρconcrete

¼ wa − wd

wa − wwð1Þ

The samples were dried at 40 and 105°C to estimate the capillaryporosity and the total (capillary and gel) porosity respectively.

Gas Permeability. A gas permeability test was conducted onconcrete cores (d ¼ 150 mm, h ¼ 50 mm) according to the recom-mendation RILEM TC 116-PCD (1999) using a CEMBUREAUpermeameter. After being saturated in water, the concrete speci-mens were dried at 40°C until a constant weight was obtained(mass change < 0.05 wt% over 24 h). Hereby the tested test spec-imens had a saturation degree of 42.0� 0.69 wt%. During themeasurement, the concrete specimens were placed in the permea-meter cells as such as the oxygen gas was prevented from leaking.The first gas pressure was set at 2 bar, and the flow rate was mea-sured using the soap bubble flow meter when a steady-state regimewas established (after about 30 min). Subsequently, the measure-ment was repeated for gas pressures of 3 and 4 bar. The measure-ments were then redone for all concrete specimens after drying at105°C (saturation degree is nihil). The gas permeability coefficient(k −m2) is calculated according to the following formula:

k ¼ 2 · Pa · Q · h · ηA · ðP2 − P2

aÞð2Þ

with Pa (101,300 N=m2) the atmospheric pressure, Q (m3=s) themeasured flow rate, h (0.05 m) the specimen height, A(π · ð0.075 mÞ2) the cross-sectional area, η (2.02 · 10−5 N · s=m2)the dynamic viscosity of oxygen gas, and P (N=m2) the appliedpressure.

Carbonation. For an accelerated carbonation test, concretecubes (100 × 100 × 100 mm3) were placed in a climate chamber(20� 2°C, RH60� 5%, 10% CO2) after 26 days of curing. Fivesurfaces of each concrete cube were coated to ensure a unidirec-tional CO2 ingress. The carbonation ingress was measured withphenolphthalein after every 28 days exposure to the high CO2

concentration.Chloride Ingress. The resistance against the ingress of chlor-

ides was evaluated by a CTH test, described in NT Build 492

Table 3. Produced Concrete Compositions

Concrete raw material CS-0 QCS-20 QCS-40 SCS-20 SCS-40

River sand 0/4 (kg=m3) 762.3 383.2 0 786.3 836.6QCS (kg=m3) — 403.1 848.3 — —SCS (kg=m3) — — — 403.1 857.7Gravel 2/8 (kg=m3) 449.9 464.0 488.3 311.9 169.9Gravel 8/16 (kg=m3) 741.7 765.0 805.0 514.1 280.1CEM I 52.5 N (kg=m3) 340 340 340 340 340Water (kg=m3) 153 153 153 153 153Superplasticizer(ml=kg binder)

5 2 2 2 2

Fig. 2. Particle size distributions of the used materials determined by laser diffraction (a) or sieving (b)

© ASCE 04014177-3 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

(1999). An external electrical potential applied axially across thespecimen (drilled core with d ¼ 100 mm h ¼ 50 mm) forceschloride ions outside to migrate into the specimen. After a certaintest duration (mostly 24 h), a silver nitrate solution is sprayed onto freshly split sections, and the chloride penetration depth(xd–mm) is measured. From these results a nonsteady-state migra-tion coefficient (D − 10−12 m2=s) can be calculated:

D¼ 0.0239 · ð273þTÞ ·LðU−2Þ · t

"xd−0.0238

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffið273þTÞ ·L · xd

U−2

r #ð3Þ

with UðVÞ the absolute value of the applied voltage, T (°C) theaverage value of the initial and final temperatures in the anolytesolution, L (mm) the thickness of the specimen, and t (h) the testduration.

Freeze-Thaw Attack with Deicing Agents. The resistanceagainst freeze-thaw attack with deicing agents was tested accordingto appendix D of NBN EN 1339 (NBN 2003). A concrete testspecimen (drilled core d ¼ 100 mm h ¼ 50 mm) with a salt sol-ution (NaCl, 3%) on top was subjected to 28 freeze-thaw cycles.After every 7 cycles, the scaled material was collected, dried for atleast 24 h at 105� 5°C, and weighed.

Statistical AnalysisThe software package SPSS Statistics was used for statistical analy-sis of the results. Analysis of variance (ANOVA) was used for com-parison of the different mixtures. The Tukey and Dunnett methodwere applied for post hoc testing when equal variances are respec-tively assumed or not assumed. A significance level of 0.05 wasused for all tests.

Results

Copper Slag as Cement Replacement

Isothermal Calorimetry of Cement PastesIn Fig. 3, the heat production rate and cumulative heat productionper gram of powder of cement pastes with varying mQCS contentare presented. It is seen that both the maximum of the heatproduction rate and the cumulative heat production decrease withincreasing mQCS content. It can also be observed that the maxi-mum of the heat production rate moves further in time when more

CEM I 52.5 N is replaced by mQCS. To estimate the possible con-tribution of the mQCS to the heat production, the heat productionrate and cumulative heat production per gram of CEM I 52.5 N arepresented in Fig. 4. It is seen that the maximum values of the heatproduction rate are similar for all mixes, but the maximum movesfurther in time with increasing mQCS content. It is also observedthat the cumulative hydration heat of the cement pastes with mQCSincrease more compared with the reference mixture without mQCS.

Compressive Strength of MortarsThe compressive strength of mortars with increasing mQCS con-tent to replace CEM I 52.5 N are presented in Table 4. Although thecompressive strength test after 7 days indicated that the addition of20 wt% mQCS seems beneficial and results in a higher strengthresistance compared with the reference, the reference mortar isthe strongest after a curing period of 28 days. Except for the goodresult after 7 days for 20 wt% mQCS addition, it can be observedthat the compressive strength decreases with increasing mQCS con-tent, for both curing times of 7 and 28 days.

Copper Slag as Aggregate Replacement

Compressive Strength of MortarsMortars with increasing QCS content in replacement of standardsand were tested for their compressive strength, and the resultsare presented in Table 4. For all mortar samples with QCS, the com-pressive strength after 7 days of curing is comparable for the differ-ent amounts of copper slag, although some results have significantdifferences between the mean values (Table 4). Compared with thereference mortar, all results for the copper slag mixtures are signifi-cantly higher. For the tests after a curing period of 28 days, thecompressive strength results are similar for the reference mortarand the mortars with QCS, but again results might differ signifi-cantly. The mortars with 40 and 60 vol% QCS give the best resultsafter 28 days, whereas the mortar with 100 vol% QCS has thelowest strength, although not significantly different from the refer-ence mortar.

Compressive Strength of ConcreteThe results of the compressive strength tests on concrete with vary-ing copper slag contents are presented in Table 5. For the mixtureswith 20 vol% copper slag (QCS-20 or SCS-20), the results arecomparable for all curing ages (no significant differences). Forthe mixtures with 40 vol% copper slag (QCS-40 and SCS-40),

Fig. 3. Heat production rate (a); cumulative heat production (b) per gram of powder (both CEM I 52.5 N and mQCS) with varying mQCS contents

© ASCE 04014177-4 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

the results are only comparable after 28 and 84 days of curing, andthe compressive strength of SCS-40 is significantly higher after 2and 7 days of curing. Compared with the reference mixture, theresults for concrete with 20 vol% and 40 vol% copper slag are re-spectively similar or higher (not in the case of QCS-40 after 2 and7 days curing). Thus, from the compressive strength tests on con-crete it can be concluded that the use of copper slag as aggregateresults in comparable or even better results.

Durability of ConcreteThe durability of concrete with copper slag as aggregate replace-ment was studied by testing several parameters, of which theresults are presented in Table 5. Regarding durability the ingressof aggressive substances is an important issue. Important concretecharacteristics regarding the ability of these substances to penetrateinto concrete are its porosity and gas permeability. Looking at theresults of the open porosity, the effect caused by the copper slagseems to be acceptable. The capillary porosity of QCS-20 andSCS-40 is comparable with the reference mixture, whereas thecapillary porosity of SCS-20 and QCS-40 is respectively signifi-cantly higher or lower. For the total porosity, there is no significantdifference for QCS-20 and QCS-40, which is also true for QCS-40,SCS-40, and the reference mixture. Again the total porosity ofSCS-20 is significantly the highest. Nonetheless, the differenceswith the reference mixture seem minimal for all mixtures. Thegas permeability is for all mixtures with copper slag lower

compared with the reference, and this for all pressures and bothdrying temperatures. This improvement is, however, only signifi-cant for mixtures SCS-20 and SCS-40 after being dried at 105°C.After drying at 40°C, the reduction is significant for SCS-20 andSCS-40 for an applied pressure of 3 bar, for SCS-40 also at 4 bar.

Both the reference concrete and the mixtures with copper slagwere subjected to a high CO2 concentration (10%). All mixturesshowed negligible carbonation up to 84 days. For the reference con-crete and SCS-40, the first limited carbonation was observed after84 days of exposure. For the other mixtures with copper slag, lim-ited carbonation was measured after the first 28 days of carbona-tion. The ingress occurred locally, explaining the high variability.For example, where carbonation was measured after 28 and 56 daysfor QCS-20, no carbonation was measured after 84 days for thesame mixture on different samples.

The migration of chlorides into concrete was tested by a CTHtest. The chloride migration coefficient was for all mixtures withcopper slag lower compared with the reference mixture. Exceptfor QCS-20, this improvement was also significant.

Different surfaces (casted, sawn, and troweled) of most mixtureswere tested for the resistance against freeze-thaw attack with de-icing agents. The results of SCS-20 are not available as the testspecimens were destroyed because of a technical defect, and thesame concrete batch could not be tested again. The technical defectalso made it impossible to have the results after 28 cycles for SCS-40. The results presented in Table 5 clearly show that the resistanceagainst freeze-thaw attack with deicing agents is worse for concretewith copper slag aggregates compared with the reference concrete.Although the effect of testing different surfaces is rather limited forthe reference mixture, there is a higher variability for the mixtureswith copper slag. However, there is no surface type showing con-sequently a better or worse performance for the different amountsand types of copper slag addition.

Discussion

Copper Slag as Cement Replacement

From the results obtained within this study, it can be concluded thatthe contribution of ground copper slag to the hydration degree of acement paste is negligible. The hydration heat reduces with increas-ing mQCS content when plotted per gram of powder because of adecreasing cement content. The increase of the hydration heat whenpresented per gram of CEM I 52.5 N is probably caused by the

Fig. 4. Heat production rate (a); cumulative heat production (b) per gram of CEM I 52.5 N with varying mQCS contents

Table 4. Average Results of the Compressive Strength Tests (N=mm2) onMortar Samples with Copper Slag as Cement (mQCS) or Sand (QCS)Replacement

Amount and typeof copper slag 7 days 28 days

Reference mortar 31.5 (0.90) 61.5 ð0.92ÞD;E20wt% mQCS 36.8 (0.51) 44.7 (0.49)40wt% mQCS 26.1 (0.17) 34.3 (0.25)60wt% mQCS 13.9 (0.10) 17.2 (0.27)20vol% QCS 54.2 ð0.66ÞA 63.4 ð0.66ÞE;F40vol% QCS 50.3 ð1.08ÞB;C 66.9 ð0.84ÞG60vol% QCS 53.3 ð0.40ÞA;B 65.3 ð0.44ÞF;G80vol% QCS 52.7 ð0.25ÞA;B 62.7 ð0.64ÞE;F100vol% QCS 48.4 ð0.70ÞC 58.2 ð0.82ÞDNote: A capital superscript indicates the groups for which the mean valuesdo not differ significantly; standard errors on the mean values are addedbetween parentheses.

© ASCE 04014177-5 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

increase of the water to cement ratio, as cement is replaced by analmost inert powder. The higher the water to cement ratio, the morewater will be available for hydration, resulting in a higher ultimatehydration degree αu, which can be calculated according to thefollowing formula (Mills 1966) cited in Baert (2009):

αu ¼1.031 · W=C0.194þW=C

ð4Þ

For all mixtures, the ultimate hydration degrees were calculatedusing their water to cement ratio and the above mentioned formula.For comparison, the hydration degree after 7 days was calculated asthe ratio of the total measured heat to the potential heat of the ce-ment. The latter was calculated using the mineralogical composi-tion of the cement (Table 2) and the potential heat release of thedifferent clinker minerals: 500, 260, 1,673, and 42 J=g for alite,belite, aluminate, and ferrite respectively. It was found that the

calculated hydration degrees after 7 days of hydration are compa-rable with the calculated ultimate hydration degrees (Fig. 5). Com-parable hydration degrees after 7 days were obtained in Baert(2009), when the effect of the water to cement ratio on the heatrelease of ordinary Portland cement pastes was studied. This con-firms the presumption that the measured increase of the hydrationheat per gram of cement can be attributed to the variation in water tocement ratio, and the effect of the copper slag is nihil. The delay ofthe main hydration peak can be explained by the set-inhibitingproperties of heavy metals in copper slag (Zain et al. 2004), seeTable 1. The reduction in compressive strength with increasing cop-per slag content is explained by the lower cement content of themixtures.

Looking into the literature, it is seen that most researchers usemuch lower replacement levels of up to 15 wt% (Al-Jabri et al.2006; Shi et al. 2008; Taha et al. 2004; Tixier et al. 1997; Zainet al. 2004). Their conclusion is that the effect of using copper slag

Table 5. Results of Compressive Strength and Durability Tests on Concrete Containing Copper Slag

Tested parameter CS-0 QCS-20 QCS-40 SCS-20 SCS-40

Fresh propertiesSlump (mm) 60 50 20 40 50Air content (%) 2.8 2.2 1.2 2.4 2.0

Compressive strength (N=mm2)2 days 34.7 (0.31)a 33.6 ð0.26ÞA;B 33.7 (0.15)a 32.2 (0.48)b 40.4 (0.25)7 days 49.3 ð0.88ÞC;D 48.2 ð0.18ÞC 51.8 ð0.78ÞD 47.9 ð0.73ÞC 56.2 (0.47)28 days 57.0 ð0.98ÞE 57.7 ð1.16ÞE 63.7 ð1.49ÞF 59.2 ð0.23ÞE 65.6 ð0.31ÞF84 days 63.9 ð1.14ÞG;H 59.9 ð1.61ÞG 69.3 ð1.61ÞH 64.0 ð1.47ÞG;H 69.0 ð3.11ÞH

Open porosity (vol%)Capillary porosity (40°C) 7.5 ð0.12ÞI 7.9 ð0.15ÞI 6.9 (0.05) 8.9 (0.20) 8.1 ð0.16ÞITotal porosity (105°C) 12.6 ð0.13ÞK 11.9 ð0.08ÞJ 12.3 ð0.08ÞJ;K 13.6 (0.14) 12.6 ð0.16ÞKGel porosity (105–40°C) 5.1 ð0.16ÞM;N 4.0 ð0.13ÞL 5.5 ð0.12ÞN 4.7 ð0.15ÞM 4.5 ð0.16ÞL;M

Gas permeability coefficient (10−17 m2)—after drying at 40°C2 bar 6.7 ð0.44ÞO 5.9 ð0.29ÞO 5.5 ð0.61ÞO 5.2 ð0.25ÞO 5.1 ð0.08ÞO3 bar 4.5 ð0.29ÞP 3.9 ð0.16ÞP;Q 3.5 ð0.38ÞP;Q 3.4 ð0.15ÞQ 3.3 ð0.07ÞQ4 bar 4.1 ð0.27ÞR 3.5 ð0.13ÞR;S 3.1 ð0.31ÞR;S 3.2 ð0.15ÞR;S 3.1 ð0.06ÞS

Gas permeability coefficient (10−17 m2)—after drying at 105°C2 bar 43.3 ð2.39ÞT 39.0 ð0.96ÞT 37.4 ð1.32ÞT 30.2 ð0.94ÞU 29.9 ð0.31ÞU3 bar 28.7 ð1.47ÞV 25.5 ð0.67ÞV 24.3 ð0.89ÞV 19.8 ð0.64ÞW 19.4 ð0.24ÞW4 bar 24.5 ð1.27ÞX 21.2 ð0.62ÞX 20.0 ð0.75ÞX 16.6 ð0.53ÞY 16.1 ð0.23ÞY

Carbonation depth (mm)28 days exposure —a 0.76 (0.17) 0.30 (0.08) 0.06 (0.04) —a

56 days exposure —a 0.12 (0.08) 0.18 (0.13) 0.12 (0.07) —a

84 days exposure 0.24 (0.09) —a 0.52 (0.16) —a 0.24 (0.08)Chloride migration coefficient (10−12 m2=s)

13.0 ð1.0ÞZ 12.5 ð1.6ÞZ 9.3 (1.1) 11.1 ð0.6ÞAA 11.1 ð0.9ÞAAScaled material due to freeze-thaw attack with deicing agents—casting surface (kg=m2)

7 cycles 0.04 (0.01) 0.53 (0.15) 0.06 (0.01) n/ab 0.02 (0.01)14 cycles 0.34 (0.10) 1.41 (0.02) 0.15 (0.04) n/ab 0.44 (0.10)21 cycles 0.70 (0.16) 2.03 (0.07) 0.42 (0.12) n/ab 1.24 (0.20)28 cycles 1.14 (0.24) 2.78 (0.14) n/ab n/ab 2.13 (0.25)

Scaled material due to freeze-thaw attack with deicing agents—sawn surface (kg=m2)7 cycles 0.17 (0.23) 0.11 (0.02) 0.24 (0.03) n/ab 0.17 (0.03)14 cycles 0.34 (0.05) 0.34 (0.03) 0.62 (0.05) n/ab 0.63 (0.06)21 cycles 0.63 (0.13) 0.90 (0.09) 1.08 (0.09) n/ab 1.23 (0.09)28 cycles 1.04 (0.23) 1.63 (0.20) n/ab n/ab 1.85 (0.16)

Scaled material due to freeze-thaw attack with deicing agents—troweled surface (kg=m2)7 cycles 0.45 (0.06) 0.07 (0.01) 0.84 (0.07) n/ab 0.65 (0.12)14 cycles 0.70 (0.11) 0.63 (0.07) 1.88 (0.08) n/ab 1.60 (0.15)21 cycles 0.85 (0.14) 1.56 (0.10) 2.54 (0.07) n/ab 2.47 (0.37)28 cycles 1.08 (0.19) 2.24 (0.14) n/ab n/ab 3.79 (0.50)

Note: A capital superscript indicates the groups for which the mean values do not differ significantly (for the results of the compressive strength, open porosity,gas permeability, and chloride migration). Standard errors on the mean values are added between parentheses.aNo carbonation observed.bBecause of a technical defect, the test specimens were destroyed, and the same mixture could not be tested again.

© ASCE 04014177-6 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

on compressive strength is limited, and only in some cases smallimprovements were noticed. Some researchers used activators suchas cement by-pass dust and/or lime, which had no effect (Al-Jabriet al. 2006) or only a limited effect (Taha et al. 2004). In Tixier et al.(1997), a positive effect on the compressive strength from addingcopper slag was noticed. This copper slag was well crystallized,and the fineness of the copper slag was comparable with theone of cement, which is not the case for the copper slag mQCSused in this study (Fig. 2). In further research, it might thus be in-teresting to start from the SCS instead of QCS, and higher fine-nesses should be aimed for.

Finally, it is worth mentioning that the replacement of cement bycopper slag results in a better resistance regarding sulphate attack(Najimi et al. 2011). The replacement with copper slag up to 15%led to more than 50% decrease in sulphate expansion, accompaniedwith less strength degradation. Microstructure studies showed thatettringite, present in the control samples, was absent in the sampleswith copper slag. The results were explained as follows (Najimiet al. 2011): a reduced C3A content because of the replacementwith copper slag, a reduction of soluble calcium hydroxide (if cop-per slag would show a mediocre pozzolanic reactivity), or a lesspermeable microstructure from the copper slag addition.

Copper Slag as Aggregate Replacement

From the compressive strength test results, it was concluded thatthe replacement of sand by copper slag had no significant effecton the strength performance of the mortars. As the compressivestrength of mortars with 40 vol%, QCS was the highest, and liter-ature suggests replacement levels of about 40–50% (Al-Jabri et al.2011, 2009a; Shi et al. 2008); the replacement levels for concretewere chosen to be 20 and 40 vol%, for both the fine fraction (withQCS) and the coarse fraction (with SCS). Regarding the compres-sive strength of this concrete compared with the reference, it wasconcluded that the results were similar or even better. Similar orslightly improved compressive strengths were expected from liter-ature (Al-Jabri et al. 2011; Shi et al. 2008; Wu et al. 2010a, b) andare related to the physical properties of the copper slag. Copperslag has a better compressibility than natural aggregates, partiallyrelieving the stress concentration, if natural sand is still the dom-inant fine aggregate holding the concrete matrix together. Also the

angular sharp edges of the copper slag can improve the cohesion ofthe concrete matrix, although the glassy surface of the copper slagcan have a negative effect on the cohesion. Furthermore copperslag as aggregate replacement can have a negative effect on thecompressive strength because of their low water absorption, leavingexcess water as discussed later in this paper.

The performance of the concrete with copper slag was comparedwith the reference concrete by means of durability indices (DI; Kaidet al. 2009). These indices are formulated as such that values higherthan 1 indicate the copper slag concrete performs better than thereference concrete. The results are presented in Fig. 6. It was foundthat the effect of aggregate replacement levels of 20 and 40 vol% waslimited regarding the open porosity and gas permeability. For SCS-20 and SCS-40, the gas permeability was even reduced comparedwith the reference. Regarding carbonation and chloride migration,the performance of the concrete with copper slag is comparable withthe one of the reference concrete. However, in the test for freeze-thawattack with deicing agents, the concretes with copper slag performedworse compared with the reference, and the surface scaling after 28standardized freeze-thaw cycles increased significantly. Dependingon the surface and the mixture the surface scaling increased by57 to 252% because of the addition of copper slag. As requiredby NBN EN 1339 (NBN 2003), the amount of scaled material shouldnot exceed the limit of 1 kg=m2. In Table 5 it is seen that even thereference mixture is slightly exceeding the limit. The performanceregarding freeze-thaw attack with deicing agents is strongly relatedto the air content of fresh concrete. As can be seen in Table 5, the aircontents are rather low for both the reference and copper slag con-crete. The European standard NBN EN 206-1 (NBN 2001b) and theBelgian standard NBN B 15-001 (NBN 2004) require an air contentof at least 4% for a nominal maximum aggregate size of 20 mm. Inthe American standard, ACI 201.2 R (ACI 2008) requires even an aircontent of 6–7%. It is thus advised to improve the resistance againstfreeze-thaw attack with deicing agents by means of, for example, airentraining agents.

An effect on strength and durability from the addition of copperslag was expected, as the low water absorption and glass-likesmooth surface create a water excess and thus a risk for bleeding(Al-Jabri et al. 2009b; Shi et al. 2008; Wu et al. 2010a) since thewater to cement ratio is kept constant. This excess water is waterthat is not absorbed by the aggregates and becomes available in thecement matrix that results in a higher content of voids, microcracks, and capillary channels, especially when the substitution rateexceeds 20% (Wu et al. 2010a). This has of course its consequencesfor the performance of the concrete. From the results reported inthis paper, it was concluded that mainly the performance regardingfreeze-thaw attack with deicing agents of the copper slag concretewas reduced compared with the reference concrete. In case ofthe copper slag concrete, it might thus not be only the air contentof the fresh concrete that resulted in the inferior performance, butalso the quality of the concrete surfaces is expected to be reducedbecause of an excess of water and thus a risk for bleeding and seg-regation (Shi et al. 2008). In order to compensate for this excess inwater, less water could be added. Al-Jabri et al. (2009b) replacedsand by copper slag keeping the workability constant and adaptingthe water to cement ratio. They concluded that a 100% replacementof the sand by copper slag results in a water reduction of 22%. Inthis study both the water to cement ratio (0.45) and amount ofsuperplasticizer (2 ml=kg binder) of the copper slag concrete werekept constant. As observed in Table 5, the slump of the copper slagmixtures is rather low, although the risk for bleeding was observed.A detailed study on the water demand of the mixtures is required.The optimum should allow a sufficient slump of the concrete, with-out the risk for bleeding.

Fig. 5. Comparison of the calculated ultimate hydration degree andhydration degree based on the measured hydration heat after 7 daysfor cement pastes with varying mQCS content

© ASCE 04014177-7 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Conclusions

Copper slag is interesting as an iron source for the production ofCRC. First of all, it is an industrial by-product that reduces the useof natural resources on the one hand and the waste production onthe other hand. Secondly, it is already used as raw material in thecement production as an iron adjusting material, and the final des-tiny of CRC is to become a raw meal for cement production. Theonly question that remained was: how to use this material withinconcrete? From literature, it was concluded that there are two pos-sibilities: as an alternative binder or as an aggregate. Both possibil-ities were tested for the copper slag available, and it was concludedthat the use of copper slag as an alternative binder was of minorinterest because no significant pozzolanic reaction was observed.The use of copper slag as aggregate replacement is, however, prom-ising. The effects on the compressive strength, open porosity, gaspermeability, resistance against carbonation, and chloride ingressseemed acceptable or even an improved performance of theconcrete was observed. The latter can be related to the physicalproperties of the copper slag, having an improved compressibilityand improved cohesion due to the angular sharp edges. Only theresistance to freeze-thaw attack with deicing agents was inferior.The performance in such an environment might however be im-proved by adapting the water content of the concrete or the additionof an air entrainer. The excess of water due to the lower water ab-sorption of the copper slag was claimed to be the main parameternegatively affecting the performance of the concrete regardingfreeze-thaw attack with deicing agents.

Acknowledgments

Financial support from the Institute for the Promotion of Innovationthrough Science and Technology in Flanders (IWT-Vlaanderen)and the Research Foundation Flanders (FWO) (Grantno. G087510 N) for this study is gratefully acknowledged.

References

ACI. (2008). “Guide to durable concrete.” 201.2 R, ACI, Farmington Hills,MI.

Al-Jabri, K. S., Al-Saidy, A. H., and Taha, R. (2011). “Effect of copper slagas a fine aggregate on the properties of cement mortars and concrete.”Constr. Build. Mater., 25(2), 933–938.

Al-Jabri, K. S., Hisada, M., Al-Oraimi, S. K., and Al-Saidy, A. H. (2009a).“Copper slag as sand replacement for high performance concrete.” Cem.Concr. Compos., 31(7), 483–488.

Al-Jabri, K. S., Hisada, M., Al-Saidy, A. H., and Al-Oraimi, S. K. (2009b).“Performance of high strength concrete made with copper slag as a fineaggregate.” Constr. Build. Mater., 23(6), 2132–2140.

Al-Jabri, K. S., Taha, R. A., Al-Hashmi, A., and Al-Harthy, A. S. (2006).“Effect of copper slag and cement by-pass dust addition on mechanicalproperties of concrete.” Constr. Build. Mater., 20(5), 322–331.

Arab, H. B. (2006). “Aggregates from construction & demolition waste inEurope.” State of play—Providing essential materials for Europe,European Aggregates Association, Brussels, Belgium.

Baert, G. (2009). “Physico-chemical interactions in Portland cement—(high volume) fly ash binders.” Ph.D. thesis, Ghent Univ., Ghent.

De Belie, N., and Robeyst, N. (2007). “Recycling of construction materi-als.” Environment-conscious construction materials and systems, Stateof the art report of TC 192-ECM, RILEM Rep. No. 37, N. Kashino, D.Van Gemert and K. Imamoto, eds., RILEM Publications S.A.R.L.,Bagneux, 11–23.

De Schepper, M., De Buysser, K., Van Driessche, I., and De Belie, N.(2013). “The regeneration of cement out of completely recyclableconcrete: Clinker production evaluation.” Constr. Build. Mater., 38,1001–1009.

De Schepper, M., Snellings, R., De Buysser, K., Van Driessche, I., andDe Belie, N. (2014). “The hydration of cement regenerated fromcompletely recyclable concrete.” Constr. Build. Mater., 60(0), 33–41.

De Schutter, G. (1999). “Hydration and temperature development ofconcrete made with blast-furnace slag cement.” Cem. Concr. Res.,29(1), 143–149.

Goni, S., Lorenzo, M. P., and Sagrera, J. L. (1994). “Durability of hydratedPortland cement with copper slag addition in NaCl+Na2SO4 medium.”Cem. Concr. Res., 24(8), 1403–1412.

0

1

2fc,2d

fc,7d

fc,28d

fc,84d

ϕ40°C

ϕ105°C

ϕ105°C-40°C

K40°C,2bar

K40°C,3barK40°C,4bar

K105°C,2bar

K105°C,3bar

K105°C,4bar

D

Scs

Sss

Sts

QCS-20QCS-40SCS-20SCS-40

Freeze thaw attack with deicing agents

Compressive strength

Porosity

Chloride diffusion

Gas permeability

Sxs = Scaled material after 28 cycles for the casting (cs), sawn (ss) or troweled (ts) surface

fc,xd = x days compressive strength

ϕx°C = open porosity after drying at x °C

D = non-steady-state chloride migration coefficient

Kx°C,nbar = gas permeability coefficient after drying at x°C with an applied pressure of n bar

111

22

000000

Fig. 6. Durability indices (dashes) to compare the performance of concrete with copper slag to the reference concrete

© ASCE 04014177-8 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Gorai, B., and Jana, R. K. (2003). “Characteristics and utilisation of copperslag—A review.” Resour. Conserv. Recycl., 39(4), 299–313.

Kaid, N., Cyr, M., Julien, S., and Khelafi, H. (2009). “Durability of con-crete containing a natural pozzolan as defined by a performance-basedapproach.” Constr. Build. Mater., 23(12), 3457–3467.

Khanzadi, M., and Behnood, A. (2009). “Mechanical properties of high-strength concrete incorporating copper slag as coarse aggregate.”Constr. Build. Mater., 23(6), 2183–2188.

Kogel, J. E., Trivedi, N. C., and Barker, J. M. (2006). Industrial mineralsand rocks, 7th Ed., Society for Mining, Metallurgy and Exploration,Englewood, CO.

Langer, W. H., Drew, L. J., and Sachs, J. S. (2004). “Aggregate and theenvironment: Production, construction, reclamation.” American Geo-logical Institute, Alexandria.

McDonough, W., and Braungart, M. (2002). Cradle to cradle: Remakingthe way we make things, 1st Ed., North Point Press, New York.

Mills, R. H. (1966). “Factors influencing cessation of hydration in water-cured cement pastes.” Symp. on the Structure of Portland Cement Pasteand Concrete, Highway Research Board Special Rep. 90, HighwayResearch Board, Washington, DC, 406–424.

Najimi, M., Sobhani, J., and Pourkhorshidi, A. R. (2011). “Durability ofcopper slag contained concrete exposed to sulfate attack.” Constr.Build. Mater., 25(4), 1895–1905.

NBN. (1976). “Testing of building materials: Resistance to freezing: Waterabsorption by capillarity.” B 05-201, NBN, Brussels, Belgium.

NBN. (2001a). “Testing hardened concrete—Part 1: Shape, dimensions andother requirements for specimens and moulds.” EN 12390-1, NBN,Brussels, Belgium.

NBN. (2001b). “Concrete—Part 1: Specification, performance, productionand conformity.” EN 206-1, NBN, Brussels, Belgium.

NBN. (2003). “Concrete paving flags—Requirements and test methods.”EN 1339, NBN, Brussels, Belgium.

NBN. (2004). “Supplement to NBN EN 206-1—concrete—specification, per-formance, production and confirmity.” B 15-001, NBN, Brussels, Belgium.

NBN. (2005). “Methods of testing cement—Part 1: Determination ofstrength.” EN 196-1, NBN, Brussels, Belgium.

NT Build 492. (1999). Concrete, mortar and cement-based repair materi-als: Chloride migration coefficient from non-steady state migrationexperiments.

RILEM TC 116-PCD. (1999). “Permeability of concrete as a criterion of itsdurability.” Final report: Concrete durability—An approach towardsperformance testing Vol. 32, RILEM Publications SARL, Bagneux,France, 163–173.

Shi, C., Meyer, C., and Behnood, A. (2008). “Utilization of copper slag incement and concrete.” Resour. Conserv. Recycl., 52(10), 1115–1120.

Taha, R., Al-Rawas, A., Al-Jabri, K., Al-Harthy, A., Hassan, H., andAl-Oraimi, S. (2004). “An overview of waste materials recycling inthe Sultanate of Oman.” Resour. Conserv. Recycl., 41(4), 293–306.

Taha, R. A., Alnuaimi, A. S., Al-Jabri, K. S., and Al-Harthy, A. S. (2007).“Evaluation of controlled low strength materials containing industrialby-products.” Build. Environ., 42(9), 3366–3372.

Tamura, M., Noguchi, T., and Tomosawa, F. (2004). “Cementitious waste-free-type completely recyclable concrete.” Proc., RILEM Int. Symp. onEnvironment-Conscious Materials and Systems for Sustainable Devel-opment, N. Kashino and Y. Ohama, eds., RILEM Publications SARL,Bagneux, France, 61–71.

Thomas, B. S., Anoop, S., and Kumar, V. S. (2012). “Utilization ofsolid waste particles as aggregates in concrete.” Procedia Eng., 38,3789–3796.

Tixier, R., Devaguptapu, R., and Mobasher, B. (1997). “The effect of cop-per slag on the hydration and mechanical properties of cementitiousmixtures.” Cem. Concr. Res., 27(10), 1569–1580.

Wu, W., Zhang, W., and Ma, G. (2010a). “Mechanical properties of copperslag reinforced concrete under dynamic compression.” Constr. Build.Mater., 24(6), 910–917.

Wu, W., Zhang, W., and Ma, G. (2010b). “Optimum content of copperslag as a fine aggregate in high strength concrete.” Mater. Des., 31(6),2878–2883.

Zain, M. F. M., Islam, M. N., Radin, S. S., and Yap, S. G. (2004). “Cement-based solidification for the safe disposal of blasted copper slag.” Cem.Concr. Compos., 26(7), 845–851.

© ASCE 04014177-9 J. Mater. Civ. Eng.

J. Mater. Civ. Eng. 2015.27.

Dow

nloa

ded

from

asc

elib

rary

.org

by

Uni

vers

iteit

Gen

t on

05/0

8/15

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.